Soft-Masked Semi-Dual Optimal Transport for Partial Domain Adaptation

2025-05-03

AI Summary

This paper addresses Partial Domain Adaptation (PDA), where the source domain’s class set strictly contains the target domain’s class set, and unsupervised target classification must be performed under domain shift. We propose a Soft-Masked Semi-Dual Optimal Transport (SM-SDOT) framework: (i) a confidence-guided soft masking mechanism for the transport matrix, enabling class-weighted source reconstruction; and (ii) a generalizable Kantorovich potential neural network that explicitly aligns class-conditional distributions. The method integrates entropy-regularized optimal transport, the semi-dual formulation, neural approximation of Kantorovich potentials, and end-to-end alternating optimization—jointly estimating class weights and soft masks. Evaluated on four standard benchmarks, SM-SDOT significantly outperforms state-of-the-art methods, demonstrating superior robustness to category mismatch and high-accuracy cross-domain transfer.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Mixed Discrete/Continuous Search
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Privacy-enhancing technologies

Abstract

Visual domain adaptation aims to learn discriminative and domain-invariant representation for an unlabeled target domain by leveraging knowledge from a labeled source domain. Partial domain adaptation (PDA) is a general and practical scenario in which the target label space is a subset of the source one. The challenges of PDA exist due to not only domain shift but also the non-identical label spaces of domains. In this paper, a Soft-masked Semi-dual Optimal Transport (SSOT) method is proposed to deal with the PDA problem. Specifically, the class weights of domains are estimated, and then a reweighed source domain is constructed, which is favorable in conducting class-conditional distribution matching with the target domain. A soft-masked transport distance matrix is constructed by category predictions, which will enhance the class-oriented representation ability of optimal transport in the shared feature space. To deal with large-scale optimal transport problems, the semi-dual formulation of the entropy-regularized Kantorovich problem is employed since it can be optimized by gradient-based algorithms. Further, a neural network is exploited to approximate the Kantorovich potential due to its strong fitting ability. This network parametrization also allows the generalization of the dual variable outside the supports of the input distribution. The SSOT model is built upon neural networks, which can be optimized alternately in an end-to-end manner. Extensive experiments are conducted on four benchmark datasets to demonstrate the effectiveness of SSOT.

Problem

• Addresses partial domain adaptation with non-identical label spaces
• Estimates class weights to enable class-conditional distribution matching
• Uses semi-dual optimal transport for scalable gradient-based optimization

Innovation

• Soft-masked transport distance matrix enhances class representation
• Semi-dual formulation optimizes large-scale transport problems
• Neural network approximates Kantorovich potential for generalization